On 19/02/2014 16:03, Daniel Murphy wrote:
"Bruno Medeiros" wrote in message news:le2i5p$28g5$1...@digitalmars.com...
[snip]
I know that a lot of work needs to be done before the frontend is ready
to be efficiently used as a library, but that doesn't make it impossible.
But anyways, those are secondary points actually. What I think is
really crucial, is that we don't have a main-stream compiler in D.
When DMD gets ported to D, and the main development of DMD happens
there, in the D version, then I'll believe a reasonable
compiler-as-library could happen.
I think you may have missed something! This is coming in the near future.
I have been working on automatically converting the DMD fronted to D
since dconf (and before). I've produced working compilers on win32,
linux32, and linux64, with the other platforms currently lacking only
because I don't have a box set up for them.
The only outstanding work for dmd is fixing layout and
commenting/formatting issues in the generated source, which I am slowly
getting through.
I expect any issues encountered with the other platforms to be minor,
and hopefully switching gdc and ldc will be fairly painless too.
I'm hoping 2.066 will be the first D-based dmd release and work on the
C++ source will cease immediately after that.
I'm glad to hear that, that sounds very promising. Especially since
dogfooding the compiler has great importance that goes beyond just the
compiler-as-a-library goal.
I think DMD will get there eventually, but... you think it will come as
soon as the DMD 2.066 release already?? As in, the official DMD, not a
fork. Walter and Andrei have said they are on board with that?
Also, Ian raises a good point, what will happen to GDC and LDC once that
switch happens? They are both very important projects to be left hanging
in the rain.
--
Bruno Medeiros
https://twitter.com/brunodomedeiros
